Multi-objective design optimisation and pareto front visualisation of radial-flux eddy current coupling for a wind generator drivetrain

Abstract

In this paper the design optimisation of a 2.2 kW, double rotor, radial axis eddy current coupling operating at 97 % efficiency, designed for a wind turbine drive train application is described. A computationally efficient finite element analysis in conjunction with gradient (MMFD) and population (NSGA) based design optimisation algorithms is used in order to obtain an optimal coupling design. The two optimisation algorithms are evaluated in terms of speed and accuracy. The comparison between copper and aluminium conductor materials revealed that the aluminium is the preferred material in terms of mass and cost. A gradient-based analysis method with regard to the input parameters of the genetic optimisation algorithm's pareto curve is proposed.